Switching 3V3

[Rajinder1268]

Hi all,
I have an application that requires 3V3 and 5V supply.
I have two PCB's and need to switch the 3V3 to the second PCB via a connector.

What is the best way of doing this please?
I have considered using opto MOSFET (relay), PNP transistor or p channel FET that is driven from a GPIO via a npn transistor - the p channel fet source is tied to 3V3. The gate tied to the source pin via 10K resistor i.e. keeping the p channel off. The gate is also tied to the collector of the NPN, so when this is switched on the p-channel gate is low and the device is switched. I am looking to switch 3V3 @ 500mA. I need the 3V3 to power a silicon labs BLE module.

Thanks in advance.

 

[stenzer]

Hi,

a simple block diagram would be helpful to show how your boards are planned to be connected, as well as to see where the 3.3 V are comming from (created at board #1 ?). Also a simple circuitry would be better, than explain it by text.

I am looking to switch 3V3 @ 500mA.

Your textual explaination sounds good. Use a P-Cahnnel enhencement MOSFET with a low R_on (P_loss = 500 mA • R_on) and check its VGS threshold voltage, by means check if the MOSFET is fully turned on when VGS is -3.3 V.

BR

 

[KlausST]

Hi,

a simple block diagram would be helpful to show how your boards are planned to be connected

I fully agree.
Without this information we can not say
* whether the MOSFET internal diode may cause problems,
* what drive voltage is available,
* whether the inrush current is too high
* ... and so on

Klaus

 

[cupoftea]

Yes , a NPN to take down the gate of a PFET sounds good..make it a logic level pfet....also, be sure to take into account the sudden inrush to the 3v3, that may take down the upstream 3v3 rail, and poss cause problems....eg reset a micro in the upstream bit...if this is the case, then you will need some inrush consideration cct aswell.......eg add a source resistor which increases vgate at the switch time...also, you can switch it on "bit by bit style"...kind of pwm it then turn it fully on...so the inrush is less

 

[Rajinder1268]

Hi,
I have attached a block diagram. I need some help in selecting a suitable p channel FET. I did think of in rush current but was not sure how to deal with it.
Any advice would be appreciated.

 

[stenzer]

Hi,

please link also the BLE module you are planning to use, to check its current consumption.

A first comment regarding your block diagram, connect a shunt resistor from the base of your NPN transistor to ground (~10 k). This ensure to keep the BJT turned-off / non-conductive during powering up of your circuitry.

BR

 

[cupoftea]

Here attached are several inrush options...you can vary them as you please...its in the free LTspice simulator....you can download free.
To "see" the current in a wire....hover over the wire and go ALT + LEFTCLICK

 

[Rajinder1268]

Hi,
The Bluetooth module is from Silicon Labs:
EFR32BG24
It's around 160mA when transmitting.
Thanks

--- Updated May 2, 2022 ---

Hi,

please link also the BLE module you are planning to use, to check its current consumption.

A first comment regarding your block diagram, connect a shunt resistor from the base of your NPN transistor to ground (~10 k). This ensure to keep the BJT turned-off / non-conductive during powering up of your circuitry.

BR

Will do, I totally missed this.

 

[d123]

Hi,

Excuse my ignorance... Why is the NPN base resistor 1k for a 10k on the PMOS? What am I missing/not understanding?

BJT base: 3.3V - 0.7V = 2.6V
2.6V/1,000 Ohms = 2.6mA
PMOS pull-up: even the complete 3.3V/10,000 Ohms = 330uA



I would have expected a 10k on the NPN base (and as mentioned, an 100k or even a 1M to ground) for a 1k on the MOSFET gate pull-up; or, using the PMOS pull-up value, an 100k on the NPN base (and 1M to ground) for the 10k PMOS pull-up.

And that's a very good MOSFET with no voltage drop across the source-drain, 3.3V in = 3.3V out... .

Genuinely curious to know reason why for 1k base resistor. Mistake or method?

 

[Rajinder1268]

Hi,

Excuse my ignorance... Why is the NPN base resistor 1k for a 10k on the PMOS? What am I missing/not understanding?

BJT base: 3.3V - 0.7V = 2.6V
2.6V/1,000 Ohms = 2.6mA
PMOS pull-up: even the complete 3.3V/10,000 Ohms = 330uA



I would have expected a 10k on the NPN base (and as mentioned, an 100k or even a 1M to ground) for a 1k on the MOSFET gate pull-up; or, using the PMOS pull-up value, an 100k on the NPN base (and 1M to ground) for the 10k PMOS pull-up.

And that's a very good MOSFET with no voltage drop across the source-drain, 3.3V in = 3.3V out... .

Genuinely curious to know reason why for 1k base resistor. Mistake or method?

Sorry it was a mistake. I was drawing by hand as an illustration..

--- Updated May 3, 2022 ---

Sorry it was a mistake. I was drawing by hand as an illustration..

Please can you show me how you got the calculations based on the resistors. Why should I have a 100k on the NPN base (and 1M to ground) for the 10k PMOS pull-up. This is a genuine question. TIA.



Excuse my ignorance... Why is the NPN base resistor 1k for a 10k on the PMOS? What am I missing/not understanding?

BJT base: 3.3V - 0.7V = 2.6V
2.6V/1,000 Ohms = 2.6mA
PMOS pull-up: even the complete 3.3V/10,000 Ohms = 330uA



I would have expected a 10k on the NPN base (and as mentioned, an 100k or even a 1M to ground) for a 1k on the MOSFET gate pull-up; or, using the PMOS pull-up value, an 100k on the NPN base (and 1M to ground) for the 10k PMOS pull-up.

And that's a very good MOSFET with no voltage drop across the source-drain, 3.3V in = 3.3V out... .

Genuinely curious to know reason why for 1k base resistor. Mistake or method?
[/QUOTE]

--- Updated May 3, 2022 ---

Please can you show me how you got the calculations based on the resistors. Why should I have a 100k on the NPN base (and 1M to ground) for the 10k PMOS pull-up. This is a genuine question. TIA.

 

[d123]

Hi,

Thanks for your explanation, I wouldn't apologize for a little mistake, my circuit designs are usually plagued with more perturbing mistakes than a typo .

As far as I remember and understand BJT resistor selection, it's typical BJT tutorial stuff, not the same as using the datasheet to select the right value based on expected hFE; it's the lazy/beginner's approach:

Presumably, we go for a BJT hFE of 10..., so if pull-up is 10k, ib should be 0.1 x ic, so Rb = 100k. If Rb is 100k, then the lazy-man's calculator says use a 1M to ground because 1M/1.1M = 0.9, and 0.9 x 2.6V (3.3V - 0.7V of the Vbe) = 2.34V means we don't lose as much of the useful/gratuitous 2.6V left over to turn the BJT on than say if we selected Rb = 100k, R pull-down = 100k, so 100k/200k = 0.5 x 2.6V = 1.3V.

Frankly, as you know, that's much of a muchness, so long as input available for Vbe is at least 0.7V (or whatever the datasheet says Vbe fully on is) and ib is around 0.1 x ic. If it were a precision circuit of some kind or the datasheet showed that hFE = 100 or 200 or whatever, I'd be more clinical about ib:ic ratio...

But I'm certain you know all this, so I'll shut up there.